QuestDB. It is built for finance workloads (can be also used for other timeseries data, like energy, or aerospace, but has heavy optimizations for common finance data patterns), it is very performant, it has been in used for years at large finance entities, and it is Apache 2.0.
Full disclosure: I am a Developer Advocate at QuestDB.
The Open Source edition does not limit any commercial use or the size of the machine you can install, as per the Apache 2.0 license terms.
If you want more enterprise-y options, like single sign on, or RBAC, there is an Enterprise edition. But Open Source is as performant as the Enterprise version. Enterprise offers also things like replication and TLS on all endpoints, which can be somehow replicated in Open Source with manual sharding or proxies.
QuestDB is just a DB. What people get wrong about KDB is that it's so much more. It's a programming language with DB capabilities. You can use for real time streaming, in memory DB and on disk DB. You can build your entire analytics on top of that. There's nothing else out there that let's you build and entire framework/platform with just one single tech stack.
Full disclosure: I am a Developer Advocate at QuestDB.
The Open Source edition does not limit any commercial use or the size of the machine you can install, as per the Apache 2.0 license terms.
If you want more enterprise-y options, like single sign on, or RBAC, there is an Enterprise edition. But Open Source is as performant as the Enterprise version. Enterprise offers also things like replication and TLS on all endpoints, which can be somehow replicated in Open Source with manual sharding or proxies.